In the 3D video encoding or multi-view application, depth images may be used to provide a more flexible viewing angle. The depth information is different from the texture information, i.e., it is not directly used for a display, but converted into a disparity vector and then used for generating a synthetic viewpoint together with the texture information.
During the generation of the synthetic viewpoint, the distortion of the depth block has very little influence on the distortion of the synthetic viewpoint under the following two conditions: 1. the distortion of the depth value does not cause the distortion of the disparity vector; and 2. the vicinity area of the texture value corresponding to the current depth value is very smooth.
It can be seen that under the above two conditions, the residual of the encoding depth block needs to occupy a code rate without improving the quality of the synthetic viewpoint.